alt Hacker News> 12 hour ahead prediction
As the comments on that article rightfully point out, restricting the data analysed to 12 hours before the resolution feels like cherry picking.
> 12 hour ahead prediction
As the comments on that article rightfully point out, restricting the data analysed to 12 hours before the resolution feels like cherry picking.
It may be cherry-picking, but I think some commenters misunderstand this (or maybe I do).
The implication seems to be "12 hours before the resolution things are obvious anyway". But if that were the case, then I could pick some wager that is obviously true but has, for example, 70% chance, and putting my money on that. If it was true that "12 hours before the resolution it's obvious what the result is", everything would be in 0% or 100% buckets. I believe getting event with 30% confidence right exactly 30% times is impressive no matter if that's 12h or 120h before.
Disclaimer: I don't know much about prediction markets, just what I understood from the blog post.